nvidia-AI.summerschool17-no.../teaching-lw/000-introduction.ipynb

495 lines
306 KiB
Text
Raw Permalink Normal View History

2017-09-22 17:26:46 +00:00
{
"metadata": {
"name": "000-introduction"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Title: Lernen und Wahrnehmen (89-7002)\n",
"Author: Thomas M. Breuel\n",
"Institution: University of Kaiserslautern"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"\n",
"from IPython.core.display import Image\n",
"def fig(x): return Image(filename=\"Figures/\"+x+\".png\")"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 2
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Course Mechanics"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Home Page)\n",
"\n",
"The course home page is at:\n",
"\n",
"http://lw.iupr.com/"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Lecture Notes)\n",
"\n",
"There is a complete set of lecture notes on the home page, as well as all notebooks and slides.\n",
"\n",
"(NB: all materials may get updated during the course)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(practical work)\n",
"\n",
"You must do practical exercises in Python; you cannot pass the course just learning the theory.\n",
"\n",
"Exercises are not \"programming exercises\": focus on interactive exploration and experimentation, don't treat them like programming assignments.\n",
"\n",
"You can download most notebooks used in lecture.\n",
"\n",
"Most exercises have fairly short solutions.\n",
"\n",
"Set up iPython Notebooks right away and become comfortable with Python and notebooks."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(background)\n",
"\n",
"- know how to program in an OOL and quickly be able to pick up Python.\n",
"- know some basic algorithms\n",
"- mathematics: derivatives and integrals, transcendental functions,linear algebra, analytic geometry, complex numbers"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(exercises)\n",
"\n",
"FIND EXERCISE TIME NOW"
]
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Perception"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(perception)\n",
"\n",
"The five senses:\n",
"\n",
"- seeing\n",
"- hearing\n",
"- feeling\n",
"- tasting\n",
"- smelling"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(extended senses)\n",
"\n",
"In engineering apps, we tend to prefer sensors that work like human senses because they are easiest to design for and easiest to understand.\n",
"\n",
"- UV, IR, X-rays\n",
"- infrasound, ultrasound\n",
"- extended chemical sensors\n",
"- particles etc.\n",
"- holographic imaging\n",
"- quantum sensing"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(common principles and techniques)\n",
"\n",
"- linear systems theory (noise, linear filters, etc.)\n",
"- statistics\n",
"- Bayesian decision theory\n",
"- dynamical systems\n",
"- multisensor integration, 3D modeling\n",
"- feedback control\n",
"- mathematical morphology\n",
"- machine learning"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(image processing)\n",
"\n",
"In this course, we focus on _image processing_ because you can learn most of the techniques you need for all different kinds of sensors.\n",
"\n",
"Imaging is easier than other sensor modalities in that you can look at the signal.\n",
"\n",
"Imaging is harder than other sensor modalities in that the signals are 2D (2D algorithms are a little harder than 1D algorithms)."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(image processing)\n",
"\n",
"Cameras have become almost universal sensors:\n",
"\n",
"- digital photography, social networking\n",
"- web cams and video conferencing\n",
"- document scanning\n",
"- bar code reading\n",
"- camera-based translation and assistive devices\n",
"- manufacturing (quality control, robot control)\n",
"- biomedical applications (diagnosis, monitoring, robotics, imaging, behavior)\n",
"- navigation\n",
"- forensics and surveillance"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(user-facing applications)\n",
"\n",
"- Google Books\n",
"- Google Streetview\n",
"- Google Glass\n",
"- Google Translate\n",
"- Bundesbahn Tickets\n",
"- Facebook face recognition\n",
"- Samsung face detection, eye control\n",
"\n",
"...\n",
"\n",
"- actually, just making your images look good\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(what does image processing do?)\n",
"\n",
"- fundamental data type: rank-2 or rank-3 array\n",
"- compression\n",
"- enhancement\n",
"- restoration\n",
"- measurements\n",
"- understanding"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(this course)\n",
"\n",
"- linear filters, gradient filters, anisotropic filters\n",
"- frequency domain methods and Fourier transforms\n",
"- template matching\n",
"- nonlinear filters and mathematical morphology\n",
"- distance transform and labeling"
]
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Machine Learning"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(approaches to image processing)\n",
"\n",
"- manual, heuristic construction\n",
"- principled, model-based construction\n",
"- machine learning"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(machine learning)\n",
"\n",
"You need:\n",
"\n",
"- large number of examples of inputs plus desired outputs\n",
"- a learning algorithm\n",
"- lots of compute power\n",
"\n",
"You get:\n",
"\n",
"- an automatically constructed \"program\" that approximates the input/output relationship\n",
"- a characterization of the expected error rate"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(other names)\n",
"\n",
"Machine learning is central to modern computer science, so people have invented lots of names for it to make it seem more related to what they are working on:\n",
"\n",
"- machine learning\n",
"- pattern recognition\n",
"- data mining\n",
"- \"big data\"\n",
"- analytics\n",
"- parametric and non-parametric statistics\n",
"- ...\n",
"\n",
"All of these use a core set of common techniques, but terminology and applications differ."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(machine learning on images)\n",
"\n",
"- image classification (image to class)\n",
"- image segmentation (image to regions)\n",
"- object recognition (parts-of-image to classes)\n",
"- image processing (image to image)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(kinds of machine learning algorithms)\n",
"\n",
"Machine learning can be based on a number of approaches:\n",
"\n",
"- statistical models of data generation (generative)\n",
"- mathematical models of decision boundaries (discriminative)\n",
"- mathematical models of neurons (neural networks)\n",
"- logical rules (expert systems, Bayesian networks)\n",
"- past examples and memory (case based reasoning)\n",
"\n",
"Despite the different approaches, they often give similar performance."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(this course)\n",
"\n",
"- nearest neighbor classifiers\n",
"- perceptrons\n",
"- linear least squares, logistic regression\n",
"- nonlinear classifiers\n",
"- generative classifiers (Gaussians), decision regions\n",
"- unsupervised learning (k-means, PCA)"
]
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Python"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(About Python)\n",
"\n",
"Dynamic, object-oriented programming language.\n",
"\n",
"Python is widely used for scripting, desktop apps, web apps, scientific and numerics, graphics, and games.\n",
"\n",
"Companies making extensive use of Python: Google, Microsoft Research, Yahoo Maps/Groups, Blender, Civ 4, Industrial Light and Magic, Walt Disney Animation, NASA...\n",
"\n",
"Multiple implementations: Python, IronPython (Python-CLR, CE), Cython, PyPy (JIT), jython (Python-JVM)"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# Python example: quicksort\n",
"from random import choice,sample\n",
"\n",
"def quicksort(l):\n",
" if len(l)<2: return l\n",
" pivot = choice(l)\n",
" lo = [x for x in l if x<pivot]\n",
" eq = [x for x in l if x==pivot]\n",
" hi = [x for x in l if x>pivot]\n",
" return quicksort(lo)+eq+quicksort(hi)\n",
"\n"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 22
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"l = sample(range(1000),10)\n",
"print l\n",
"print quicksort(l)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"[280, 466, 29, 773, 976, 948, 932, 402, 603, 2]\n",
"[2, 29, 280, 402, 466, 603, 773, 932, 948, 976]\n"
]
}
],
"prompt_number": 23
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Python is multi-paradigm)\n",
"\n",
"- procedural\n",
"- functional\n",
"- object-oriented\n",
"- array-oriented"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Python properties)\n",
"\n",
"- interactive\n",
"- dynamically typed\n",
"- duck typed\n",
"- reflective\n",
"- interpreted (usually)\n",
"- indentation is significant"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"## iPython Notebooks\n",
"fig(\"ipython-notebook\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAIXCAIAAADPCrcyAAAACXBIWXMAAAsTAAALEwEAmpwYAAAA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\nDDNZQ26iFmcLeY2CCFGjGVkm
"prompt_number": 11,
"text": [
"<IPython.core.display.Image at 0x454d0d0>"
]
}
],
"prompt_number": 11
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"fig(\"tiobe\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"png": "iVBORw0KGgoAAAANSUhEUgAAAoAAAAHjCAIAAAA8JzkdAAAAA3NCSVQICAjb4U/gAAAgAElEQVR4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\nj2QpItuJrOWuXdrm8RO8vWDN
"prompt_number": 4,
"text": [
"<IPython.core.display.Image at 0x4541dd0>"
]
}
],
"prompt_number": 4
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Learning Python)\n",
"\n",
"- quick introduction in exercises\n",
"- lots of free materials on the web\n",
"- lots of books (e.g., 150 books on Amazon; cf. 626 for Java, 50 for Ruby)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Important Software)\n",
"\n",
"- Python 2.x\n",
"- Ubuntu (easiest to install)\n",
"- NumPy\n",
"- SciPy\n",
"- Matplotlib\n",
"- scikit.learn\n",
"- gimp\n",
"- UNIX shell scripts"
]
}
],
"metadata": {}
}
]
}